Method and apparatus for protecting markup language document against cross-site scripting attack

摘要

A method for decomposing a web application into one or more domain sandboxes ensures that the contents of each sandbox are protected from attacks on the web application outside that sandbox. Sandboxing is achieved on a per-element basis by identifying content that should be put under protection, generating a secure domain name for the identified content, and replacing the identified content with a unique reference (e.g., an iframe) to the generated secure domain. The identified content is then served only from the generated secure domain.